DBN-DNN: discrimination and classification of VNG sequence using deep neural network framework in the EMD domain. Issue 6 (1st November 2020)

• Record Type: Journal Article
• Authors: Ben Slama, Amine; Sahli, Hanene; Mouelhi, Aymen; Marrakchi, Jihene; Sayadi, Mounir; Trabelsi, Hedi
• Is Part Of: Computer methods in biomechanics and biomedical engineering. Volume 8:Issue 6(2020)
• Page Start: 681
• Page End: 690
• Publication Date: 2020-11-01
• Subjects: Eye movement -- (VNG) system -- vestibular disorder -- empirical mode decomposition -- deep neural network; Imaging systems in biology -- Periodicals; Imaging systems in medicine -- Periodicals; Biomechanics -- Data processing -- Periodicals; Biomedical engineering -- Periodicals; 616.0757
• DOI: 10.1080/21681163.2020.1799867
• Languages: English
• ISSNs: 2168-1163

Abstract

ABSTRACT: The Vestibulo-ocular response VOR is characterized by a smooth pursuit eye movements in one direction, called slow phase of ocular nystagmus, interrupted by resetting saccades fast phase of nystagmus in the other direction. Recording of ocular nystagmus during vestibular tests does not quantify the true response of the vestibulo-ocular reflex (VOR). In order to extract the real VOR, our study is focused on nystagmus analysis using videonystagmography (VNG) technique based on measuring amplitude vibration of eyeball movement. The effectiveness of this attendance is severely topic to the attention and the experience of ENT doctors. In this case, automatic methods of image analysis offer the possibility of obtaining a homogeneous, objective and above all fast diagnosis of vestibular disorder. In this paper, a fully automatic system based on nystagmus parameter analysis using a pupil detection algorithm is proposed. After a segmentation stage, a deep neural Network based classification method is applied on 90 eye movement recordings from videonystagmography (VNG) containing two types of peripheral vestibular disorders and normal patients. Experimental results obtained after several simulation, show the efficiency of the proposed methodology when compared with other classification methods.